Cerebras CEO Andrew Feldman speaks to the media at the Colovore office in Santa Clara, Calif., on March 12, 2024.
The Washington Post | Getty Images
Cerebras CEO Andrew Feldman admitted that his artificial intelligence chipmaker made a mistake last week when it didn’t immediately explain its decision to withdraw its registration for an IPO.
In a LinkedIn post late Sunday, Feldman wrote that the company still wants to go public but has changed significantly since its initial filing a year ago. The company wants to revise parts of its prospectus before selling shares to the public.
“Given that the business has improved in meaningful ways we decided to withdraw so that we can re-file with updated financials, strategy information including our approach to this the [sic] rapidly changing AI landscape,” Feldman wrote.
Days before filing its withdrawal notice on Friday, Cerebras announced a $1.1 billion funding round at a valuation of $ 8.1 billion. Some of the investors in the new round, including Tiger Global and 1789 Capital, where Donald Trump Jr. is a partner, weren’t named in the 2024 filing, he added.
“We made this call because it’s in the best interest of our investors, partners, and team — and it will allow potential investors to better understand the value of the business when we enter the public markets,” Feldman wrote, without providing a timeline for a new filing.
In its prospectus, Cerebras characterized itself as a company that produces large-scale chips for training and running AI models. This year the company has added cloud business as it operates data centers that can handle incoming requests from AI models.
What’s remained is Cerebras’ insistence that its hardware outperforms graphics processing units (GPUs), a market that Nvidia dominates but where Advanced Micro Devices is trying to play catchup. AMD said on Monday that OpenAI committed to setting up to 6 gigawatts’ worth of the company’s AI processors and could end up owning 10% of the chipmaker.
WATCH: Cerebras CEO: Here’s why our chips are a more efficient alternative to Nvidia
